Weather prediction using image data with two different model architectures: a basic neural network and ResNet50. This project leverages a dataset of weather images to train models that can predict weather conditions from images.
The dataset used for this project can be downloaded from Kaggle. It contains a folder named "dataset" which has multiple subfolders with images corresponding to different weather conditions.
Download the pre-trained models from the links below:
The weather2.ipynb notebook is used for training a model with a basic neural network architecture. Running this notebook will generate the sequential.h5 file (trained model). Alternatively, you can directly download the trained model from the link provided above.
The resnet_model_train.ipynb notebook is used for training a model with the ResNet50 architecture. The pre-trained model can be downloaded from the link provided above.
Note: Go through the code in the notebooks and comment/uncomment lines as per your need.
To host the weather prediction model on your local system, use the modelfile.py script located in the ./file upload/ directory. This script will take any image as input and predict the weather condition.
- Download the dataset from the Kaggle link provided above.
- Use the
weather2.ipynbandresnet_model_train.ipynbnotebooks to train the models or download the pre-trained models from the provided links. - To run the prediction model locally, execute the
modelfile.pyscript.
Ensure you have the required libraries installed as specified in the notebooks.
Feel free to customize the code as needed to fit your requirements. Happy predicting!